Galactic abundance gradients the metalrich end

Given the various systematic effects on massive-star abundances, some of which have been discussed above, it is clear that to investigate systematic differences between populations it is preferable to control the sample such that these

Figure 6.2. The oxygen-abundance gradient in the Milky Way from B-type stars; squares are points of Smartt & Rolleston (1997), asterisks the high-Galactic-latitude stars of Smartt et al. (2001b). Oxygen abundance is plotted as a function of the Galactocentric radial distance. Note the apparent flattening of the gradient within the solar circle (Rg ~ 8.5). The solid line is a fit to the stellar points of Smartt and Rolleston and is in good agreement with the gradient obtained from H ii regions (~—0.7 dex kpc-1).

Figure 6.2. The oxygen-abundance gradient in the Milky Way from B-type stars; squares are points of Smartt & Rolleston (1997), asterisks the high-Galactic-latitude stars of Smartt et al. (2001b). Oxygen abundance is plotted as a function of the Galactocentric radial distance. Note the apparent flattening of the gradient within the solar circle (Rg ~ 8.5). The solid line is a fit to the stellar points of Smartt and Rolleston and is in good agreement with the gradient obtained from H ii regions (~—0.7 dex kpc-1).

systematics are minimized. Smartt & Rolleston (1997) and Rolleston et al. (2000) attempted this in their study of the abundance gradient of the Milky Way using B-type stars. Their work, though based on LTE methods, was well controlled such that the differential abundances between objects were not greatly influenced by NLTE effects, differences in effective temperature or choice of absorption lines. They found good agreement with the abundance gradients derived from H ii regions but crucially, from the point of view of this conference, did not sample the metal-rich end of the abundance gradient as it is extrapolated towards the Galactic Centre, similarly to the results of Daflon & Cunha (2004). The reason of course is the increasing extinction in this region. However, they attempted to rectify this (Smartt et al. 2001b) by observing four B-type stars at Galactic latitudes between 5 and 8 degrees above the Galactic plane, and Galactic longitudes between -40 and +20 degrees. Surprisingly all four objects, lying between 3 and 5 kpc from the Galactic Centre, turn out to have [O/H] similar to those of nearby B-type stars, rather than the expected supersolar [O/H] as predicted from the abundance gradient (Figure 6.2).

The interpretation of the results is somewhat unclear, however, since the abundances of magnesium and silicon are consistent with the expected abundance gradient, while the origin of these high-latitude stars is also uncertain.

The local-group galaxy M31 presents us with an excellent alternative possibility in which to study metal-rich massive stars; extinction is low and the generally accepted abundance gradient as derived from H11 regions implies supersolar [O/H] in the inner regions of this galaxy. Its remoteness, however, constrains us to the study of supergiant stars, and A-type and B-type supergiants have been used for this purpose. Venn et al. (2000) used A-type supergiants to probe the iron abundance directly, also deriving [O/H]. However, early B-type supergiants are ideal tracers of [O/H] since they have many O 11 lines in the optical that one can use for determining the oxygen abundance, which is of special interest here because it can be used to compare directly with nebular results. The first attempts at this were those of Smartt et al. (2001a), one star, and Trundle et al. (2002), an additional six stars.

In general we find the surprising result that supergiant stars display a rather flat gradient for galactocentric distances between 5 and 25 kpc, whereas the H ii results exhibit a strong gradient in this region. Smartt et al. (2001a) also demonstrated that the nebular results, which are based on calibrations of the R23 index, depend very sensitively on which calibration is used in the metal-rich domain (their object lying in the inner region of M31). They give the example of a nebula coincident with the position of the association (OB10) containing their star whose R23 ratio implies [O/H] values of 9.07 dex using Zaritsky et al. (1994) or 8.90 dex using McGaugh (1991) compared with their LTE abundance of 8.69 dex. Trundle et al. (2002) developed this theme further, showing that the calibration of Pilyugin (2001a, 2001b) implied even lower values of [O/H] at galactocentric distances of 5 kpc. Subsequently the stellar results have been improved by the application of NLTE methods including the effects of line-blanketing and winds (see elsewhere in these proceedings for a more complete description). The results are summarized in Figure 6.3, where one can see that, of the five stars analysed, three have subsolar [O/H] while two stars are mildly oxygen-rich. Furthermore, the two oxygen-rich stars belong to the same association, OB78, suggesting that this may be a characteristic of the association. In any case, it is clear that we have again failed to discover any massive star that could be said to be 'metal-rich'.

Was this article helpful?

0 0

Post a comment